Targeting RNA-binding protein ADAR1

Utilizing Receptor.AI’s pocket and hit ID workflows to
target nuclear protein in immunological diseases

2

novel allosteric pockets
targeted

1.4M

focused library
screened

4

potent hits out of
204 candidates

18x

interferon induction by
lead-like compound

*Pocket 1 anf Pocket 2 identified by Receptor.AI

01/ Background

  • The goal is to design an allosteric inhibitor for RNA-binding protein ADAR1 and avoid off-target effects with ADAR2.
  • Only a few known allosteric inhibitors exist.
  • Known allosteric pockets are poorly druggable.

02/ Methodology

  • 2 novel allosteric pockets identified by
    Receptor.AI's proprietary pocket detection AI model.
  • Virtual screening performed for 1.4 M focused library.
  • 1000 ranked compounds prioritized.
  • 209 compounds subjected to in vitro validation after
    AI-guided hit candidates selection.
  • Experimental validation performed using a
    high throughput p110 knockout cell-based assay.
  • Hit compounds confirmed by the dose-response analysis.

03/ Results

  • Criteria for a hit compound was established as a
    5x fold increase in interferon induction compared to the control.
  • Desirable outcome to surpass the efficacy of siRNA alternatives was a 10x fold increase.
  • 4 hit compounds with interferon inducing activities were identified.
  • Lead-like compound with 18x fold interferon induction.
*Hit compounds identified
*Interferon induction observed at 25 uM for Receptor.AI hit compounds 1-4 obtained through virtual screening, a competitor small molecule, and ADAR1 siRNA
*Dose-Response relationship for Compound 1,
Compound 2 and Competitor Compound
  • 2 hits exhibit comparable or superior maximal interferon induction with lower EC50 in comparison to a competing compound.
  • This was achieved on a 2.5x smaller screening library
    (209 against 500 for competitors).
  • Active scaffolds have been selected for further
    series expansion and optimization.